Inspiration
As a student who has to watch a lot of videos, it does get quite overwhelming. I often zone out when watching videos, due to the complicated terminology or how long the video is. My friends are also on the same boat as me, because they also find it hard to be able to focus on a 30 minute lecture without having their thoughts wander elsewhere. This then got me to think, how can someone be efficient at taking notes without this challenge? Simple! AI Noted. AI Noted helps speed up the process, by having the app do the hard work for you, while you get clear, personalized notes to improve your quality of learning.
What it does
AI Noted is an app that generates notes based on an audio file upload from a user. Equipped with Google's Speech to Text API, the app transcribes the video. Then, the user is prompted to select what format of notes they want, between:
- Cornell Notes
- Outline Notes
- Mind Map
- Charting Method
- Sentence Method
The user is also prompted to select the level of depth of the program, with the left side representing basic, key points, and the right side representing as much detail as possible. There is also a text box where the user can input the purpose of the notes, which allows the program to tailor it to the user's needs.
How we built it
We used Google's Speech to Text API in order to retrieve the transcription of the video provided by the user. We then created another endpoint in our program that generated the notes using Google Gemini's 2.5 Flash model, provided with the transcription, the depth, and the context behind the notes.
Challenges we ran into
Although we completed the project, it was no easy task. Configuring the API from Google Cloud was difficult as there were many steps to set up the API key and prevent it from getting leaked. Furthermore, deploying the project onto Github was not easy as I just recently learned how to upload it on to my local machine using Git, and then pushing it onto the remote repository on Github.
Accomplishments that we're proud of
Although we ran into many challenges, we were proud of many things. Firstly, I am proud of the fact that this was the first project that I had made in the year 2026. I also sharpened many of my coding skills, including Node.js, Google Cloud, etc.
What we learned
We also learned many concepts along the way, like the concepts of Git. I always thought that Github was about uploading coding files, and the Git on my local machine was complex. However, when I learned how to use Git, it was easy, and also allowed for me to be able to seamlessly deploy my program on to Github. This will set me up for the next hackathon I am doing, where I would be in a team and would have to utilize the local machine Git to collaborate easily. I also learned ESM Node.js. I used to use CommonJS in order to make some of my other projects, but this is the first time when I used ESM to configure my project. I made the switch when I realized that ESM was more modern than CommonJS.
What's next for AI, Noted
Here are the short term, medium term, and long term improvements that I am looking forward to do for AI, Noted:
Short term
- Improve the UX of the buttons of the program
- Ensure less wait-time when transcribing audio
- Incorporate functionality to convert other audio files, not just a .wav
Medium Term
- Make a login system for notes to be saved on the internet, without the user having to download it.
- Add basic accessibility features like font changes and high contrast mode
- Deploy this app on a website through tools like Netlify.
Long Term
- Allow users to directly copy the link off of YouTube, without having to downlaod and the reupload it on my app
- Add more complex accessibility tools
- Add more features that expand the use of this app, including providing study guides, practice questions, or utilizing past videos uploaded in order to get a better idea of how the user's lecture is like.
Log in or sign up for Devpost to join the conversation.